What problem does this paper attempt to address?

The key problem that this paper attempts to solve is: **How to use physical descriptors through machine - learning methods to predict the micro - chemical composition and creep strength of superalloys, and be able to perform effective extrapolation prediction in cases involving new elements**. ### Specific problems: 1. **Limitations of traditional methods**: The traditional CALPHAD method relies on element - specific thermodynamic models, which limits its wide application in new materials. In addition, existing machine - learning models usually directly use alloy compositions as descriptors, and this method performs poorly when dealing with data containing new elements. 2. **Prediction of the micro - chemical composition of multiphase alloys**: The performance of superalloys depends on their two - phase microstructures (matrix and precipitate phases). Accurately predicting the relative phase fractions and phase compositions is a crucial step in predicting alloy performance. 3. **Creep strength prediction**: The creep strength of superalloys at high temperatures is crucial for their applications. It is necessary to develop a model that can effectively predict the creep rupture life of single - crystal nickel - based superalloys. ### Solutions: 1. **A set of element - independent physical descriptors are proposed**: These descriptors are based on the distribution of atomic - electronic properties of alloys, enabling the model to be applied to any material, even those containing previously unconsidered elements. 2. **Use Gaussian Process Regression (GPR)**: Combined with the Automatic Relevance Determination (ARD) kernel function, it ensures that the model selects the most relevant features and can better fit the data. 3. **A probability correction method is developed**: It is used to correct the prediction results of phase compositions to ensure that physical constraints are met (such as the sum of total concentration and total phase fraction being 1, and the minimum material loss assumption). 4. **Heat - treatment descriptors are introduced**: According to the Ostwald ripening theory, new heat - treatment descriptors are designed to capture the microstructural evolution of alloys under different heat - treatment conditions. 5. **Extension to creep strength prediction**: By combining the microstructure prediction model and metallurgical knowledge, a more accurate creep rupture life prediction model is developed. ### Main contributions: - Element - independent physical descriptors are proposed, enabling the model to perform effective extrapolation on data containing new elements. - A probability correction method is developed, which improves the accuracy of phase composition prediction. - It is proved that the proposed model has good generalization ability in multiple superalloy systems, especially in high - entropy superalloys (HESA) and new superalloys containing Re/Ru. - By introducing heat - treatment descriptors and metallurgical knowledge, the accuracy of creep rupture life prediction is significantly improved. These improvements provide a powerful tool for accelerating the design and development of superalloys.

Machine learning superalloy microchemistry and creep strength from physical descriptors

An explainable machine learning model for superalloys creep life prediction coupling with physical metallurgy models and CALPHAD

Modern data analytics approach to predict creep of high-temperature alloys

Composition, heat treatment, microstructure and loading condition based machine learning prediction of creep life of superalloys

Evolution analysis of γ' precipitate coarsening in Co-based superalloys using kinetic theory and machine learning

Predictive Modeling of High-Entropy Alloys and Amorphous Metallic Alloys Using Machine Learning

Quantitative models of high temperature creep microstructure-property correlation of a nickel-based single crystal superalloy with physical and statistical features

Using Machine Learning and Feature Engineering to Characterize Limited Material Datasets of High-Entropy Alloys

Coupling Physics in Machine Learning to Predict Properties of High-temperatures Alloys

Machine Learning for Alloy Composition and Process Optimization

Machine learning-assisted determination of material chemical compositions: a study case on Ni-base superalloy

Predicting creep rupture life of Ni-based single crystal superalloys using divide-and-conquer approach based machine learning

Mapping the creep life of nickel-based SX superalloys in a large compositional space by a two-model linkage machine learning method

Physical metallurgy-guided machine learning and artificial intelligent design of ultrahigh-strength stainless steel

A modified divide-and-conquer based machine learning method for predicting creep life of superalloys

Deep learning accelerates the development of Ni-based single crystal superalloys: A physical-constrained neural network for creep rupture life prediction

Machine learning-based predictions and analyses of the creep rupture life of the Ni-based single crystal superalloy

Multi-Scale Crystal Plasticity Model of Creep Responses in Nickel-Based Superalloys

High-throughput map design of creep life in low-alloy steels by integrating machine learning with a genetic algorithm

Alloy Design for Mechanical Properties: Conquering the Length Scales

Machine learning metallic glass critical cooling rates through elemental and molecular simulation based featurization